Stephen J. Turnbull wrote:
It's me FKtPp ;) writes:
> Stephen J. Turnbull wrote:
> > Moving to XEmacs Patches for discussion.
> >
> > It's me FKtPp ;) writes:
> >
> > > I've created one with my two mail pasted in.
> >
> > > What about this attached patch?
> >
> > Much better. Does it do what you need it to do?
> >
>
>
> Sorry, I have to say NO.
I don't understand what it's not doing. Please explain what it's
supposed to do in the tests you've presented. Also eventually you're
going to need to explain in the context of nXML whether it does the
right thing there, too.
Sorry. The difference is in the reassignment of those two BEG and END
variable. They were used to determind the extent/overlay
starting/ending point, and in XEmacs these can not be a value outside
buffer or it will cause the "Argment out of range" error.
So in my first copy of overlay.el I make them limited in range
[(point-min), (point-max)], but after sending out the email I findout
this could break functionality if the buffer is narrowed.
Then I composed the second path, whth BEG and END limited in range [0,
(buffer-size)], but, again, I find this even worse:
XEmacs don't accept 0 as extend/overlay starting point, and report
"Argment out of range"
(buffer-size) is 1 character less than (point-max) in a widen buffer.
The second copy is totally broken... :'( sorry for no testing
After some rethought I change the range to [1, (1+ (buffer-size))] , and
run some code assumed would be failed if using the official XEmacs
overlay.el.
That's the whole story, wish it is clear enough for you to understand
the situation.
Thanks
FktPp
_______________________________________________
XEmacs-Patches mailing list
XEmacs-Patches(a)xemacs.org
http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-patches